Tektronix TBS2072, Руководство пользователя, Страница: 80 / 142

Analyzing a waveform
About FFT windows
The FFT algorithm applies a ‘window’ process to the source waveform record to ‘shape’ the record so that the start and stop
values for the FFT waveform are close to the same amplitude. Starting and stopping the waveform at close to the same
amplitude reduces adding arti
fi cial waveform that are not present in the actual signal. Using a window on the source signal
results in an FFT waveform that more accurately represents the source signal frequency components.
The different window shapes are trade-offs between frequency accuracy and magnitude accuracy. What you want to
measure, and your source signal characteristics, help you to select which window to use. Use the following guidelines to
select the best window for your signal analysis needs. The Hanning window is a good starting point for FFT measurements.
Table 2: FFT windows
Window type Window ‘shape’
Hanning
This is a very good window for measuring amplitude accuracy but less so for resolving
frequencies.
Use Hanning for measuring sine, periodic, and narrow band random noise. This window
works on transients or bursts where the signal levels before and after the event are
signi
fi cantly different.
Rectangular
This is the best type of window for resolving frequencies that are very close to the same
value but worst for accurately measuring the amplitude of those frequencies. It is the best
type for measuring the frequency spectrum of nonrepetitive signals and measuring frequency
components near DC.
Use Rectangular for measuring transients or bursts where the signal level before and after the
event are almost equal. Also, use this window for equal-amplitude sine waves with frequencies
that are very close, and for broadband random noise with a relatively slow varying spectrum.
62 TBS2000 Series User Manual